Go G, H kI.a K,a L L / D = submergence/diameter ratio N o , phase, dimensionless y Z, Z, rr = total pressure, atm. = column diameter, in. or ft. = absorption efficiency = (yly,)/y, = superficial gas mass velocity, lb./hr.-sq.ft. = superficial gas molar velocity, 1b.-moles/hr.-sq.ft. = Henry's law constant, atm./mole fraction = volumetric liquid phase mass transfer coefficient, = volumetric overall mass transfer coefficient based = nongas flow submergence, in. or ft. of liquid = number of overall transfer units based on gas = mole fraction of oxygen in air = observed height of aerated mass, in. or ft. = equivalent height of clear liquid (also L ) , in. or 1b.-moles/hr.-cu,ft.-mole fraction on gas phase, lb. -moles/hr.-cu. f t.-atm.ft.The decomposition of iron pentacarbonyl a t pressures between one and 400 Torr was studied from 120" to 300°C. Up to about 200°C. the rate was primarily limited by surface kinetics of adsorbed carbonyl. Gas phase diffusional resistances were controlling a t higher temperatures. I n the kinetically controlled region the deposition rate was correlated by a Langmuir type of equation. Corrections for diffusion resistances in the kinetic region and kinetic resistances in the diffusion region were found to be appreciable.

The decomposition of nickel tetracarbonyl a t pressures between 20 and 200 torrs was studied a t temperotures from 100" to 225°C. In the cross-flow system, the rate in the absence of appreciable dilution was primarily limited by kinetic resistances a t temperatures up to about 175°C. over an effective Reynolds number range of 5 to 100. Gas phase diffusion01 resistances were controlling a t higher temperatures. Corrections for natural convection were employed in both the kinetic and diffusional regions of control.

The use of the Chilton-Colburn analogy to obtain an effective film thickness over which the Stefan-Maxwell equations can be integrated was confirmed for binary mixtures and found opplicoble for ternary, quaternary, and presumably higher order mixtures under conditions of nonequimolal counterdiffusion in a differential convective flow system. A method of averaging binary diffusivities was developed for use in computing the Schmidt number in these multicomponent systems. It was also found necessary to modify the usual Shenvood number, if there was a large molecular weight c h n g e across the diffusion barrier when the Chilton-Colburn analogy was employed.One of the most dacult current problems in chemical engineering involves mass transfer in multicomponent gaseous mixtures under conditions of convective flow. For many years the concept of a film resistance to diffusion, originally proposed by Noyes and Whitney ( 1 ) and subsequently extended by Nernst ( 2 ) , was sufficient for most applications involving systems with two components or one diffusing component in a mixture of nondiffusing components ( 3 ) . Whitman ( 4 ) proposed the additive resistances of these films, and Fischbeck (5) and Tu et al. (6) extended their utility to interfaces at which a chemical reaction is occurrin However, Bischoff and Froment (7) reaction and kinetic resistances for real systems without the use of computational equipment. In recent years the penetration theory of Higbie ( 8 ) , with suitable modifications, has been gaining in popularity (9 to 1 2 ) , and Sherwood et al. have recently suggested turbulent film (13, 1 4 ) and parallel eddy and molecular diffusion ( 1 5 ) models. However, the film concept is still widely used industrially and is quite satisfactory for many applications.The use of analogies between momentum, heat, and mass transfer is especially he1 ful in defining a resistance ject is by Sherwood ( 1 6 ) . However, if these analogies are to be employed with the kinetic theory of gases to solve multicomponent diffusion problems, a model such,as eff ective f i l m , surface renewal, etc., must be employed. In addition, multicomponent systems present problems, because the analogies between momentum and heat transfer no longer bear a direct similarity with mass transfer. The diffusion rates of various components in a mixture are not necessarily directly related to the heat and momentum transport rates. For instance, the diffusion rates of a given component, or of the mixture as a whole, can vary in both sign and magnitude under conditions where the heat or momentum transfer is relatively invariant. Ackerman (1 7) first reported this difficult and showed how total heat ponent, as well as by the transport parameters of the bulk stream. Mickley et al.( 1 8 ) extended this concept to the case where the flow to or from an interface was essentially independent of bulk stream conditions. Such a situation can occur, for instance, in transpiration or ablative coolhave pointed out t f e difficulty of treating simultaneous to ma...

In a current research and development program a coal-fired atmospheric fluidized bed combustor is being designed to supply the heat to a closed cycle gas turbine cogeneration system. The major technical effort is directed towards the design of the in-bed heat exchanger, which is required to operate near bed temperature. This high temperature (850° C) exposes the heat exchanger tubes to potentially severe sulfidation. The corrosion behavior depends upon the intimate details of the bed environment and may be related to the occurrence of localized areas of low oxygen partial pressure and high sulfur partial pressure. This paper describes a series of measurements of oxygen partial pressure at various locations within a fluidized bed. The bed, containing densely packed heat exchanger tubes, was operated under various conditions to observe the effect of coal mixing and devolatilization on local oxygen activity. Substantial variations of oxygen partial pressure (below 10−14 atmospheres) were observed. It was noted that these locally severe variations could be substantially modified by changes in coal mixing (as through coal port design). The experiments with varying coal size suggest that rapid devolatilization is desirable and would reduce the extent of locally corrosive environments.

No abstract